1. Field of the Invention
The present invention relates to vacuum gauges such as Bayard-Alpert (BA) ionization gauges and to systems and methods for operating and calibrating such gauges.
2. Discussion of the Prior Art and Summary of the Invention
The BA gauge is the simplest known non-magnetic means of measuring very low pressures and has been widely used worldwide essentially unchanged since being disclosed in U.S. Pat. No. 2,605,431 in 1952.
A BA gauge system consists of a BA gauge and controller circuitry. A prior art BA gauge consists of a heated cathode for emitting electrons, a cylindrical grid or anode for accelerating the emitted electrons to ionizing energy, a very small cross section collector electrode on axis for collecting the ions formed within the anode volume by collision of energetic electrons with gas molecules and a vacuum envelope surrounding the gauge electrodes and attaching to a vacuum system wherein an unknown gas pressure is to be measured. Controller circuitry consists of a means to apply suitable potentials to the anode, to the cathode and to the collector electrode, means for heating the cathode to provide a controlled electron emission current and means for measuring the collector current and means for calculating and displaying the indicated pressure.
To a first approximation, the ion current, i.sub.+, to the collector electrode is proportional to the electron emission current i.sub.-, and to the gas density in the gauge or at constant temperature to the gas pressure, P.sub.G, in the gauge. Thus, EQU i.sub.+ =Si.sub.- P.sub.G Eq. 1
where S is a constant of proportionality commonly called the gauge sensitivity.
S can be calculated by measuring i.sub.+ and i.sub.- when the gas pressure P.sub.GK in the gauge can be determined by other calibration means. Thus, EQU S=i.sub.+ /(i.sub.- P.sub.GK) Eq. 2 PA1 1. Calibrate the ionization gauge system in situ utilizing, for example, an expensive spinning rotor gauge. Such a solution is being actively promoted now by several spinning rotor gauge suppliers. The prior art gauge may change calibration at anytime, so frequent in situ calibration is required. This alternative provides the best accuracy but is extremely time consuming, inconvenient and costly. PA1 2. Send the ionization gauge system to a separate calibration facility and have the system calibrated--an expensive, time consuming alternative at best because the prior art gauge may change calibration anytime with use. This alternative enables the user to comply with certain governmental requirements but is unlikely to result in improved accuracy over the long term. PA1 3. Use the generic calibration data supplied by the manufacturer assuming it is accurate for the prior art BA gauge in use. This generic calibration data is typically obtained by measurements on one or more prototypes and consists of a fixed value of the sensitivity S for a given gas type and a list of the nominal applied electrode potentials.
where P.sub.GK is the known pressure in the gauge. The value of S thus determined can be utilized to calculate the value of an unknown gauge pressure P.sub.x which produces an ion current, i.sub.+x at the same value of i.sub.- used to determine S in Eq. 2 provided conditions in the gauge do not change. Thus, EQU P.sub.X =i.sub.+X /(i.sub.- S) Eq. 3
Very complex and costly non-BA gauges have been described which are claimed to provide excellent accuracy of measurement over a limited pressure range. However, these laboratory devices are entirely unsuitable for everyday use in research and industry and have yielded few clues if any as to how to improve the accuracy of the simple prior art BA gauge.
Numerous studies over the years have demonstrated conclusively that S is not constant in prior art BA gauges. For example, Poulter and Sutton at the National Physical Laboratory in the United Kingdom reported that a typical prior art BA gauge calibration "drifted at a rate of -1.4% per 100 operating hours when it was kept under vacuum but similar gauges exposed to atmospheric conditions showed sharp changes of up to 25% in their sensitivity", K. F. Poulter & C. M. Sutton, Vacuum, 31, 147-150 (1981). In another representative study, Tilford at the U.S. National Bureau of Standards reported that typical prior art nude BA gauges (BA gauges without vacuum envelopes) had sensitivities which varied from 70% to 110% of that specified by the manufacturer, C. R. Tilford, J. Vac. Sci. Technol., A1(2), 152-162 (1983). In another study, repeated calibrations of seven prior art "broad range" BA gauges showed sensitivities varying from 52% to 67% of their specified values, C. R. Tilford, K. E. McCullogh, H. S. Woong, J. Vac. Sci. Technol., 20(4), 1140-1143 (1982).
Ionization gauge systems must be calibrated against a primary or secondary pressure standard to be useful. Users of prior art BA gauge systems have been provided with three alternatives to secure required accuracy of vacuum measurement.
The huge majority of users opt for the last alternative because of the large expense and inconvenience of either of the other alternatives even though using generic calibration data provides vacuum measurements grossly in error.
What is needed is better means for providing generic calibration data that will yield more accurate vacuum measurements.
Many researchers have pointed out for years that S is not a constant but depends on gas species, electron energies, emission current, electric field distributions, gas density, kinetic energy of the gas molecules in the gauge, plus several other parameters. But from all this work there have been no solid clues on how to improve the accuracy of the prior art BA gauge and the inaccuracy remains the same as when the device was invented. From the vantage point achieved through computer simulation of electron and ion trajectories in a BA gauge, it is now possible to discern deficiencies in the prior art which help cause the observed large inaccuracies in prior art BA gauge systems.
It is well known that the performance of BA ionization gauges can be seriously affected by a phenomenon known as surface ion desorption. An effective way of minimizing this effect is by utilizing an anode which has minimum surface area. Thus, to minimize surface ionization inaccuracies all low pressure BA gauge designs utilize transparent grids with minimal surface area. When such open grids are used, energetic electrons are not confined to the anode volume but travel a significant fraction of their total path length outside the grid. These energetic electrons traveling outside the anode volume can impinge on exposed insulating surfaces and uncontrollably change the surface potential of the exposed insulator. Their trajectories may also be changed by uncontrolled potentials outside the anode volume. Utilizing computer simulation, one can determine if energetic charged particles will impact an exposed surface or if trajectories are influenced by potentials outside the anode volume for any given configuration of surfaces, potentials, energy and initial trajectory of the charged particles.
The total electric charge distribution in a BA gauge is the sum of the charge distribution on the surfaces exposed to impact by charged particles and the charge distribution due to free charges within the gauge volume. In prior art BA gauges, means have not been provided for adequately fixing the electric charge distribution in that region of the gauge accessible to energetic charged particles. Thus, the electric charge distribution can vary from measurement-to-measurement in the same gauge or from gauge-to-gauge at any given pressure. Thus, because the electric charge distribution which exists during calibration cannot be duplicated during use, inaccurate pressure indications result.
Applicants have found by computer simulation that seemingly trivial changes in the electrode geometry or in surface potentials in the gauge volume accessible to energetic charged particles can cause large changes in electron trajectories. It is well known that changes in electron trajectories cause changes in the radial position, r.sub.o, of ion formation resulting in changes in the angular momentum, m r.sub.o v.sub.T, of the ion, where m is the mass of the ion and v.sub.T is the tangential component of its velocity. Changes in angular momentum cause the probability of ion collection to change as is well known, thus causing the gauge sensitivity to change. Changes in gauge sensitivity cause measurement inaccuracies.
Prior art BA gauges can be categorized as to the manner in which and the degree to which the electric charge distribution on surfaces exposed to charged particle impact in the gauge are uncontrolled.